Abstract

Ischemic preconditioning (IP) protects the rat liver. In pigs, in which hepatic tolerance to ischemia is similar to that in humans, information on IP is lacking. Therefore, in enflurane-anesthetized pigs, hepatic vessels were occluded for 120 min (protocol 1) or 200 min (protocol 2) without (control) and with IP (3 times 10 min ischemia-reperfusion each). In protocol 1, cumulative bile flow (CBF) during reperfusion was greater in IP (47.3 ± 5.2 ml/8 h) than in control (17.1 ± 7.8 ml/8 h,P < 0.05). ATP content tended to recover toward normal during reperfusion in IP, whereas it remained at ischemic levels in control. Serum enzyme concentrations increased similarly during reperfusion, and <1% hepatocytes were necrotic or stained terminal deosynucleotidyl transferase-mediated dUTP nick-end labeling-positive in control and IP groups. In protocol 2, no differences in CBF, ATP, or serum enzyme concentrations during reperfusion were measured between control and IP groups, except for a somewhat reduced lactate dehydrogenase in IP. The number of necrotic or terminal deosynucleotidyl transferase-mediated dUTP nick-end labeling-positive hepatocytes tended to be greater in the IP than the control group. Thus IP provides some functional protection against reversible ischemia but no protection during prolonged ischemia in pigs.

  • ischemia
  • reperfusion
  • cumulative bile flow
  • lactate dehydrogenase ATP
  • apoptosis
  • necrosis

ischemic preconditioning is characterized by the delay of infarct development during prolonged and severe myocardial ischemia by one or more preceding short episodes of ischemia and reperfusion. Since its original description by Murry and co-workers in anesthetized dogs (30), the existence of this phenomenon has been confirmed in a variety of animal species including humans (for review, see Ref.7).

In isolated rat livers, one cycle of 10-min ischemia-reperfusion protects against 90 min of normothermic ischemia, i.e., increases bile flow and reduces enzyme release compared with nonpreconditioned controls (6). In rats in vivo, ischemic preconditioning by occlusion of the hepatic artery and portal vein attenuates irreversible tissue damage (36, 48) and enzyme release (34-36, 48), and it increases liver function (48) and survival (27) following 40–90 min of normothermic ischemia.

Mechanistic information on ischemic preconditioning obtained in rat hearts, however, was sometimes not transferable to other species (5, 10). Also, the ability of isolated hepatocytes as well as the normal liver in vivo to tolerate anoxia/ischemia is species dependent. Whereas, following 150 min of anoxia and 60 min of reoxygenation, only 30% of isolated human hepatocytes were irreversibly damaged, as indicated by trypan blue uptake, >60% of isolated rat hepatocytes demonstrated irreversible damage (4). Additionally, whereas 40 min of normothermic ischemia induces hepatocyte necrosis in rats (48), the human liver tolerates up to 85 min of normothermic ischemia without signs of irreversible damage, as assessed by the lack of enzyme release (15). In pigs, in contrast to rats, hepatic tolerance to prolonged normothermic ischemia is similar to that in humans (33). Furthermore, because the pig liver is increasingly being utilized for xenotransplantation in the pig-to-human model (21, 41), mechanisms enhancing hepatic tolerance to ischemia that is inevitably associated with organ preservation are gaining importance.

The present study therefore analyzed the consequences of ischemic preconditioning on hepatic function and morphology and the animal's survival following either 120 or 200 min of normothermic ischemia in pigs. In contrast to previous studies in rats, mesenteric congestion during clamping of the portal vein and the hepatic artery was avoided by opening a shunt between a large mesenteric and a femoral vein.

METHODS

The experimental protocols employed in this study were approved by the bioethical committee of the district of Düsseldorf, and they adhere to the guiding principles of the American Physiological Society.

Fifty-seven farm pigs of either sex (30–55 kg, mean of 38 ± 5 kg) were fasted for 12 h before surgery, initially sedated using ketamine hydrochloride (1 g im), and then anesthetized with thiopental (Trapanal, 500 mg iv). Through a midline cervical incision, the trachea was intubated for connection to a respirator (Dräger, Lübeck, Germany). Anesthesia was then maintained using enflurane (1–1.5%) with an oxygen-nitrous oxide mixture (40%:60%). Arterial blood gases were monitored frequently in the initial stages of the preparation until stable and then periodically throughout the study (Radiometer, Copenhagen, Denmark). Rectal temperature was monitored and body temperature was kept above 37°C by the use of a heated surgical table and drapes. Through the cervical incision, a carotid artery and both internal jugular veins were isolated. The artery was cannulated with a polyethylene catheter for the measurement of arterial pressure (Bell and Howell Type 4–327I, Pasadena, CA). The jugular veins were cannulated for volume replacement using warmed 0.9% saline and for the measurement of central venous pressure (CVP; Bell and Howell Type 4–327I, Pasadena, CA). One femoral vein was isolated and cannulated with a 20-Fr catheter for shunting of abdominal blood during portal vein and hepatic artery occlusion. The liver was exposed through a midline incision. The common bile duct was cannulated and drained to an unpressurized reservoir to continuously measure bile flow; the cystic duct was occluded. A large branch of a mesenteric vein was dissected and cannulated with a 18-Fr catheter, which could then be connected to the femoral vein catheter. The urinary bladder was cannulated and drained.

Hepatic Plasma Flow

A bolus of indocyanine green (ICG, 0.5 mg/kg) was used to determine hepatic plasma flow (HPF) before ischemia and after 1 h of reperfusion. After injection of ICG, 2-ml venous blood samples were taken at 3, 6, 9, 12, 15, 20, 30, 40, 50, 60, 70, 80, and 90 min. Plasma ICG concentration (mg/l) was analyzed using a spectrophotometer at 804 nm (model 8452, Hewlett-Packard, Palo Alto, CA) and plotted versus time. The area under the curve was calculated using SYSTAT software (Urbana, IL), integrating the term[ICG](t)=Aeαt+Beβt where [ICG](t) is the plasma concentration of ICG at time t, A and B are the zero-time intercepts, and α and β are the slopes of the two exponentials. The ICG extraction fraction (EF) was determined from the rate constantsk 2 and k 3 of the disappearance curve according to Grainger et al. (9):k 1 = (αA + βB)/(A + B);k 2 = αβ/k 1;k 3 = α + β − (k 1 + k 2); and finally EF =k 2/(k 2 +k 3). HPF was then calculated by dividing the total dose of ICG injected by the product of the area under the ICG-time curve times EF. A significant collateral circulation to the liver was excluded at the end of each study before KCl infusion by reocclusion of the portal vein and the hepatic artery and subsequent intravenous injection of 100 ml of methylene blue. In none of the animals did the liver, in contrast to all other organs, stain with methylene blue.

Hepatic Function

Hepatic function was quantified by continuous measurement of hepatic bile flow. In analogy to the approach proposed by Bolli et al. (3) to quantify myocardial stunning as the deficit of contractile function over time, the total bile flow deficit during reperfusion (BFD, %) was calculated.

Serum Enzyme Concentrations

Venous blood samples were taken at baseline, at the end of the preconditioning and the prolonged ischemic period, and every hour throughout the reperfusion period. Serum glutamic oxaloacetic transaminase (GOT) and serum lactate dehydrogenase (LDH) were measured using the following reactions (37, 47)GOT:L­Aspartate+2­Oxoglutarate L­Glutamate+Oxaloacetate Oxaloacetate+NADH+H+L­Malate+NAD+LDH:Pyruvate+NADH+H+Lactate+NAD+ The decrease in NADH at a wavelength of 340 nm was measured using a spectrophotometer.

Hepatic ATP Content

Three to four scalpel biopsies (∼100 mg each) from the middle liver lobe were taken and immediately stored in liquid nitrogen. Samples requiring more than 1–2 s for acquisition were not used for this analysis. The analytical procedures (HPLC) have been described in detail previously (13).

Hepatic Glycogen and Lactate Contents

Glycogen and lactate were measured from the neutralized perchloric acid-tissue extracts. Glycogen was hydrolyzed by amyloglucosidase (5 U/ml final activity, Boehringer Mannheim) at pH 4.8 (0.2 M acetate buffer), 40°C, by being shakened for 2 h. The formed glucose was determined using an ultraviolet spectrophotometer and the hexokinase/glucose 6-phosphate dehydrogenase reaction (Glucoquant Glucose/HK, Boehringer Mannheim). Lactate was measured according to Noll (32) by dehydrogenation with NAD and lactate dehydrogenase (LDH), followed by transamination of the formed pyruvate with glutamate/glutamate pyruvate transaminase to shift the equilibrium to quantitative NADH formation. NADH was measured using an ultraviolet spectrophotometer (32).

Histology

Biopsies were taken from all liver lobes after 8 h (protocol 1) and 5 h (protocol 2) reperfusion. As long as ATP generation is maintained, either by glycolysis or by oxidative phosphorylation, oxidative stress induces apoptosis rather than necrosis in human T cells (23) and bovine endothelial cells (24). Indeed, a preserved, although at a reduced level, hepatic ATP content during 90–110 min of normothermic ischemia has been measured previously in anesthetized dogs (8), and therefore both the extent of necrosis and the number of terminal deosynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL)-positive hepatocytes, as an estimate of apoptosis, were quantified.

The specimens were fixed in Formalin, embedded in paraffin, and stained with hematoxylin-eosin. Sections of each specimen were examined using light microscopy (DMSL, Leica, Bernsheim, Germany) connected to a video camera module (DC 100, Leica). Up to 20 high power fields (magnification ×250) of each specimen were analyzed. The amount of necrotic hepatocytes, characterized by enhanced acidophilia and nuclear pyknosis, was counted and expressed as a percentage of the total number of hepatocytes in each tissue section, i.e., related to a total of 5,000–6,000 hepatocytes. Figure1 shows a representative micrograph.

Fig. 1.

Representative micrograph showing an area of necrotic hepatocytes, characterized by enhanced acidophilia and nuclear pyknosis.A: magnification ×200; B: magnification ×400.

DNA-strand breaks were detected in situ by TUNEL-technique (40). In detail, sections were dewaxed, rehydrated, microwave irradiated, and covered with terminal deoxynucleotidyl transferase-FITC-conjugated dUTP (Boehringer Mannheim). Sections were incubated in this solution in a humidified chamber for 45 min at 37°C. After sections were rinsed with a phosphate-balanced salt solution, they were counterstained with bisbenzimide (HOE-33342, 1 μg/ml) for 15 min to permit a quantitative comparison between nuclei with and without DNA strand breaks. Positive (pretreatment with DNAse) and negative (omitting terminal deoxynucleotidyl transferase) controls of TUNEL were included in the protocol. Slides were directly analyzed under a fluorescence microscope (Leica DMLB). The number of TUNEL-positive hepatocytes was counted and expressed as percentage of the total number of hepatocytes in each tissue section, i.e., related to a total of 4,000–5,000 hepatocytes. Figure2 shows a representative micrograph.

Fig. 2.

Representative micrograph showing terminal deosynucleotidly trasferase-mediated dUTP nick-end labeling (TUNEL)-positive hepatocytes. A: TUNEL-staining, magnification ×200;B: counterstaining with HOE-33342, magnification ×200.

All biopsies were examined by the same investigator, who did not know the assignment of specimens to a particular group of pigs.

Experimental Protocols

A scheme is given in Fig.3. Each experiment started with the measurement of HPF by intravenous injection of ICG. During the 90-min observation period, bile flow was quantified. Before occlusion of the portal vein and hepatic artery, biopsies were taken for the determination of ATP, glycogen, and lactate contents.

Fig. 3.

Scheme of the experimental protocol. In each experiment, cumulative bile flow was continously quantified. Hepatic plasma flow was measured before the portal vein and hepatic artery occlusion and at 1 h of reperfusion. Biopsies (Histo) for the determination of ATP, glycogen, and lactate contents were taken before and at the end of the portal vein and hepatic artery occlusion as well as at the end of reperfusion. In groups 2 and 4, an additional set of biopsies was taken following 3 cycles of 10-min portal vein and hepatic artery occlusion followed by 10 min of reperfusion each. ICG, indocyanine green; IP, ischemic preconditioning.

Protocol 1: 120 min of Prolonged Ischemia

After injection of 500 IU/kg heparin, the portal vein and hepatic artery were occluded, and the mesenterico-femoral shunt was opened (group 1; n = 11). Bile flow during the 120-min normothermic, no-flow ischemia was measured, and before release of the occlusion biopsies for the determination of ATP, glycogen and lactate contents were taken. Immediately upon reperfusion, the mesenterico-femoral shunt was closed. Throughout the 8-h reperfusion period, bile flow was measured. At 1 h of reperfusion, HPF was once more assessed. Before termination of the study, another set of biopsies for the determination of ATP, glycogen, and lactate contents as well as for the histological analysis was taken.

After injection of 500 IU/kg heparin, pigs underwent three cycles of 10-min portal vein and hepatic artery occlusion followed by 10-min reperfusion each (group 2; n = 11). During each occlusion period, the mesenterico-femoral shunt was opened. At the end of the three cycles of ischemia and reperfusion, biopsies for the determination of ATP, glycogen, and lactate contents were taken. Thereafter, the protocol of group 2 was identical to that ofgroup 1.

Protocol 2: 200 min of Prolonged Ischemia

The 120 min of ischemia followed by 8 h of reperfusion induced only negligible hepatic injury. Therefore, to critically test the hypothesis that ischemic preconditioning delays the development of irreversible hepatic injury, the duration of ischemia was increased to 200 min.

The protocol of group 3 (n = 21) was identical to that of group 1, except that the ischemic period was prolonged to 200 min, and reperfusion was limited to 5 h. The protocol of group 4 (n = 14) was identical to that of group 2, again except that the ischemic period was prolonged to 200 min, and reperfusion was limited to 5 h.

Data Analysis and Statistics

Hemodynamic data were recorded on an eight-channel recorder (Gould MK 200A, Cleveland, OH) and stored directly to the hard disk of an AT-type computer. Hemodynamic parameters were digitized and recorded over a 20-s period using CORDAT II software (43). Hemodynamic parameters reported are heart rate (HR), mean arterial pressure (MAP), and CVP. Calculation of all hemodynamic parameters was done on a beat-to-beat basis, and data were then averaged. Statistical analysis was performed using SYSTAT software (Urbana, IL). Data on systemic hemodynamics, HPF, as well as cumulative bile flow during the time course of the experiment in the two groups of pigs each inprotocols 1 and 2 were compared using a two-way analysis of variance for repeated measures. For the statistical analysis, only data from pigs surviving the 8-h reperfusion period (protocol 1) or the 5-h reperfusion period (protocol 2) were used.

When significant differences were detected, individual mean values were compared using least significant difference post hoc tests. All data are reported as means ± SE, and a P value <0.05 was accepted as indicating a significant difference in mean values. Because values for serum enzyme concentrations were not normally distributed, as assessed by a skewness coefficient (skewness/standard error of skewness) and a kurtosis coefficient (kurtosis/standard error of kurtosis) <2, median values ± the confidence interval are reported. The time courses of enzyme concentration in the two groups of pigs each in protocols 1 and 2 were compared using the two-sample Kolmogorov-Smirnov test. A P value <0.05 was accepted as indicating a significant difference in median values.

RESULTS

Protocol 1: 120 min of Prolonged Ischemia

Survival in protocol 1.

Three pigs of group 1 and two pigs of group 2died during early reperfusion from low cardiac output failure or ventricular fibrillation. Therefore, data were derived from eight pigs of group 1 and nine pigs of group 2.

Hemodynamics in protocol 1.

During the preconditioning ischemic periods, CVP and MAP tended to decrease, whereas HR tended to increase (data not shown). At the end of the preconditioning period, HR, MAP, and CVP were similar to their baseline values (Table 1). After 120 min of ischemia, HR was increased to a similar extent in both groups (P < 0.05), and MAP was decreased (not significant). At 8 h of reperfusion, HR was still slightly elevated (not significant) and MAP was decreased (P < 0.05).

View this table:
Table 1.

Hemodynamics, cumulative bile flow, and metabolism in control and preconditioned pigs

HPF and CBF in protocol 1.

HPF during reperfusion was reduced to a comparable extent in both groups (group 1: from 229 ± 31 to 120 ± 26 ml/min; group 2: from 241 ± 25 to 110 ± 15 ml/min, both P < 0.05). CBF was comparable between the two groups at baseline (group 1: 27.9 ± 5.2 ml/h;group 2: 29.3 ± 3.0 ml/h) and was almost completely stopped in both groups during ischemia. Whereas during the 8-h reperfusion period, CBF remained severely depressed in group 1 (BFD: 92 ± 5%), and it recovered somewhat in group 2 (BFD: 73 ± 6%, P < 0.05) (Table 1).

Serum enzyme concentrations in protocol 1.

During reperfusion, serum enzyme concentrations were increased (bothP < 0.05) without a significant difference between groups (Fig. 4).

Fig. 4.

Serum enzyme concentration during the time course of theprotocol 1 in control and IP pigs. Rep, reperfusion; Isch, ischemia. Data are median ± confidence interval; *P < 0.05 vs. baseline.

Hepatic metabolism in protocol 1.

Ischemic preconditioning decreased ATP content (not significant). At 120 min of ischemia, ATP content was, however, again similar in both groups. Whereas ATP content at 8 h of reperfusion remained severely depressed in group 1, there was a trend toward recovery in group 2 (Table 1). With the use of the data of ATP content and CBF flow during baseline, ischemia, and reperfusion, a close correlation was apparent (CBF = 13.87/ATP content − 2.18; r = 0.79, P < 0.001).

Ischemic preconditioning decreased glycogen content. At 120 min of ischemia, glycogen content was decreased to similar values in both groups, and it was decreased even further until the end of reperfusion.

Lactate content was almost unchanged by ischemic preconditioning. At 120 min of ischemia, lactate content was increased to nearly identical values in both groups, and it was decreased only slightly until the end of reperfusion.

Histology in protocol 1.

In both groups of pigs, <1% hepatocytes were necrotic. Ingroups 1 and 2, 0.6 ± 0.4% and 0.9 ± 0.7% of hepatocytes, respectively, stained positive for TUNEL following 120 min of ischemia and 8 h of reperfusion.

Protocol 2: 200 min of Prolonged Ischemia

Survival in protocol 2.

Fifteen pigs of group 3 and eight pigs of group 4died during early reperfusion from low cardiac output failure or ventricular fibrillation. Therefore, data were derived from six pigs each in group 3 and group 4.

Hemodynamics in protocol 2.

During the preconditioning ischemic periods, CVP and MAP tended to decrease, whereas HR tended to increase (data not shown). At the end of the preconditioning period, HR, MAP, and CVP were similar to the respective baseline values (Table 2). After 200 min of ischemia, HR was increased to a similar extent in both groups (P < 0.05) and MAP was decreased. During reperfusion, HR increased further in group 3(P < 0.05 vs. baseline), whereas HR was slightly reduced in group 4. MAP was decreased (P < 0.05) in both groups.

View this table:
Table 2.

Hemodynamics, cumulative bile flow, and metabolism in control and preconditioned pigs

HPF and CBF in protocol 2.

HPF during reperfusion was reduced to a comparable extent in both groups (group 3: from 253 ± 21 to 148 ± 19 ml/min; group 4: from 245 ± 21 to 123 ± 16 ml/min, both P < 0.05). CBF during baseline, ischemia, and reperfusion did not differ significantly between the two groups (Table 2). BFD was also comparable between group 3 (85 ± 8%) and group 4 (86 ± 8%).

Serum enzyme concentrations in protocol 2.

During reperfusion after the prolonged ischemic period, serum enzyme concentrations were increased (Fig. 5). There was a tendency toward somewhat attenuated serum GOT and LDH concentrations in group 4 compared with group 3, although, apart from LDH data at 5 h of reperfusion, these differences were not significant.

Fig. 5.

Serum enzyme concentration during the time course of theprotocol 2 in control and IP pigs. Data are median ± confidence interval; *P < 0.05 vs. baseline;a P < 0.05 vs. other group.

Hepatic metabolism in protocol 2.

Ischemic preconditioning reduced ATP content (not significant). At 200 min of ischemia, ATP content was decreased to a comparable extent ingroups 3 and 4, and it remained severely depressed in both groups during reperfusion (Table 2).

Ischemic preconditioning decreased glycogen content (P< 0.05). At 200 min of ischemia, glycogen content was decreased to a somewhat greater extent in group 4 (not significant). During reperfusion, glycogen content decreased further in group 3, whereas it remained unchanged in group 4.

Ischemic preconditioning slightly increased lactate content. At 200 min of ischemia, lactate content was increased to a greater extent ingroup 3 than in group 4 (P < 0.05). At the end of reperfusion, lactate content was, however, similar in both groups.

Histology in protocol 2.

In group 3, 5.0 ± 0.8% hepatocytes were necrotic following 200 min of ischemia and 5 h of reperfusion. Ingroup 4, 18.5 ± 8.9% hepatocytes were necrotic. Also, there were more TUNEL-positive hepatocytes in group 4(15.8 ± 8.9%) than in group 3 (5.0 ± 0.8%).

DISCUSSION

In the present study, ischemic preconditioning provided some functional protection following shorter (and fully reversible) periods of ischemia but did not attenuate functional impairment and irreversible damage following longer periods of ischemia.

Experimental Protocol

The preconditioning procedure with three cycles of ischemia-reperfusion was adapted from prior experiments. Preconditioning with several cycles of ischemia-reperfusion protected the heart (38, 45) and kidneys (22, 46) against the consequences of prolonged ischemia. In the present study, preconditioning by three cycles of ischemia-reperfusion did not attenuate hepatocyte necrosis and apoptosis following 200 min of ischemia. Whether a single episode of ischemia-reperfusion, as used in several studies in rats, would have protected the liver remains unclear at present.

The present study was performed in anesthetized, fully heparinized pigs to allow the use of a portocaval shunt. Therefore, the stability of the preparation was limited, which could potentially affect enzyme release and detection of necrosis and apoptosis. Plasma enzyme concentrations, however, were almost maximal within 4–7 h following prolonged periods of liver ischemia in prior studies (1, 18). Prolonging the reperfusion period might have improved the detection of necrosis and apoptosis in both experimental protocols, as demonstrated recently (12). In this study, the detectable amount of apoptotic cells was doubled when reperfusion was prolonged from 60 min to 2 or 4 days. Therefore, the absolute number of apoptotic hepatocytes might have been underestimated in the present study. Also, the amount of necrotic cells in protocol 2 might have been underestimated due to the limited reperfusion period of 5 h. Whether or not with prolongation of reperfusion to 8 h a protective effect of preconditioning might have become apparent, remains unclear at present.

HPF

The measured baseline and reperfusion values for HPF in the present study agree well with previous data in pigs (28) and rats (35). Importantly, no differences in postischemic HPF were measured between groups 1 and 2; therefore, the functional protection of preconditioned livers was not caused by an improved postischemic HPF.

CBF

After 120 min of ischemia, CBF did not recover in group 1, whereas it recovered somewhat in group 2 (Table 1). Similarly, in a previous study in anesthetized rats, recovery of CBF was accelerated in preconditioned over that in control livers (48), although, in contrast to the present study, improved postischemic CBF in that study was associated with a reduction of hepatocyte necrosis in the preconditioned livers.

CBF closely correlates to the activity of the sodium-potassium ATPase, which in turn is directly related to the intracellular ATP concentration (42). Indeed, in the present study as well as in a previous study in the anesthetized rat (48), the observed improvement in postischemic CBF in preconditioned livers was associated with increased ATP content.

Hepatic Metabolism

Baseline glycogen and lactate contents were comparable to that observed previously in fasted rats (2) and pigs (20). Glycogenolysis during prolonged ischemia, as estimated from the differences in glycogen content before and at the end of the prolonged ischemic period (Tables 1 and 2), was lower in the preconditioned than in nonpreconditioned groups. Furthermore, after 200 min of ischemia, lactate content in preconditioned livers was significantly lower than in control livers. Similar results have been obtained in the heart (31).

After 200 min of ischemia, ATP content did not recover in control and preconditioned livers. Also, after 120 min of ischemia, ATP content did not recover during the 8-h reperfusion period in controls. Similar results have been obtained in the anesthetized dog following 90–110 min of portal vein and hepatic artery occlusion (8). In contrast, ATP content recovered somewhat in preconditioned livers during 8 h of reperfusion. Recovery of energy-rich phosphates during reperfusion was also accelerated by ischemic preconditioning in the pig heart (29).

Serum Enzyme Concentrations

In the present study, heparin was given to permit use of a mesenterico-femoral shunt. Heparin, however, substantially attenuates the increases in serum enzyme concentrations following hepatic ischemia-reperfusion in pigs (25). Therefore, the effect of ischemia-reperfusion on serum enzyme concentrations per se as well as the decreases observed following ischemic preconditioning were most likely underestimated in the present study.

In rats, 40 min (48) or 90 min (27,34-36) of normothermic ischemia caused hepatocyte necrosis (36, 48) and increased enzyme release during reperfusion (27, 34-36, 48), both of which were attenuated by ischemic preconditioning. In the present study, following 120 min of no-flow ischemia, <1% hepatocytes were necrotic or apoptotic; however, a substantial increase in the serum LDH and GOT concentrations was observed during reperfusion, which was not affected by ischemic preconditioning. Also in other cell types, such as cardiomyocytes, a substantial cytosolic enzyme release occurs in the absence of irreversible cell injury, and only with severe membrane lesions enzyme release truly reflects irreversible injury (16). Serum enzyme concentrations during reperfusion were somewhat further increased with prolongation of ischemia to 200 min, an ischemic duration inducing irreversible hepatic injury. This time, serum enzyme concentrations during reperfusion tended to be reduced following ischemic preconditioning.

Taken together, these data indicate that serum enzyme concentrations cannot easily be used as surrogate markers for the extent of irreversible damage in hepatic ischemia-reperfusion injury.

Survival Rate

There was a tendency toward increased survival rate in the preconditioned (group 4) pigs compared with the placebo (group 3) pigs following 200 min of ischemia. Similarly, survival rate was increased by ischemic preconditioning in rats (27). In the present study, this reduction in the mortality rate could not be attributed to reduced irreversible tissue damage. Recently, it has been demonstrated that pigs surviving prolonged periods of ischemia (330 min) differed from those who died early during reperfusion in the plasma lactate concentration, but not in the extent of irreversible tissue injury (1). Also in the present study, lactate concentrations differed significantly between preconditioned and placebo pigs (Table 2). Therefore, increased survival might be achieved by ischemic preconditioning, independent from tissue salvage.

Hepatic Morphology

In rats (27, 34-36) and pigs (11,33), 90 min of normothermic ischemia regularly caused hepatocyte necrosis. In these studies, hepatic ischemia was induced by occlusion of both the hepatic artery and portal vein. In contrast, in a recent study in rats (39), the liver remained almost completely viable following 120 min of ischemia when mesenteric congestion during ischemia was avoided by use of a portosystemic shunt. Also in pigs, the duration of ischemia tolerated without development of irreversible tissue damage appeared to be larger when splanchnic decompression was used. In pigs with a portocaval shunt, 2 wk after a 330-min occlusion of the hepatic artery and portal vein, only 30% of hepatic parenchyma was necrotic (1). Similarly, in the present study only a small percentage of hepatocytes was necrotic following 200 min of ischemia, and necrosis was completely absent when the duration of ischemia was limited to 120 min. Thus the liver itself appears to be much more resistant to ischemia, in contrast to what is expected from results of previous studies (11, 27,33-36).

Increasing the duration of ischemia to 200 min induced hepatocyte necrosis and apoptosis in the present study. Whereas ischemic preconditioning delays the development of irreversible tissue damage in the heart (30), both the extent of necrotic and apoptotic cell death tended to be increased in preconditioned livers. Whether the positive TUNEL staining truly reflected apoptotic cell death or a severe, yet still reversible form of damage to the DNA, as recently reported for the ischemic heart (19), is probably not important for the present study; with prolonged ischemia, there clearly was this form of damage in addition to necrosis, and both of these morphological manifestations of ischemic damage were certainly not reduced with ischemic preconditioning.

This finding is in contrast to previous studies in rats (36,48), in which ischemic preconditioning by clamping of both the hepatic artery and portal vein reduced hepatocyte necrosis. Clamping the hepatic artery and portal vein might induce congestion of the intestine, and preconditioning of the intestine significantly reduced the increase in serum LDH concentration during reperfusion following 90 min of intestinal ischemia in rats (14, 44). However, because no direct comparison of protection by ischemic preconditioning in the presence or absence of splanchnic compression is available in the literature, it remains to be resolved whether the lack of hepatic protection by ischemic preconditioning in the present study refers to species differences or to the lack of splanchnic compression.

The results of the present study suggest that ischemic preconditioning does not induce protection against irreversible ischemic injury in hepatocytes. Liu and co-workers (26) were the first to suggest that either cell differentiation and/or the presence of contractile elements might be prerequiste(s) for ischemic preconditioning to protect. In their study, protection from preconditioning was absent in a cell line derived from human embryonic kidney (HEK-293) as well as in undifferentiated myoblasts, but it was present in differentiated myotubes (C2C12cells). This hypothesis is, at least with respect to the morphological outcome, supported by the results of the present investigation in the liver and by a recent in vivo study in rat kidneys also demonstrating no alteration in the morphological outcome by ischemic preconditioning (17).

Acknowledgments

The expert technical assistance of P. Gres, I. Konietzka, A. van de Sand, and U. Beste is appreciated. We are grateful to Prof. Dr. F.-W. Eigler (Dept. of General Surgery), Prof. Dr. H. Goebell (Dept. of Gastroenterology) and Prof. Dr. U. Schmidt (Dept. of Pathology) who supported these studies and helped prepare the paper.

Footnotes

  • Address for reprint requests and other correspondence: R. Schulz, Dept. of Pathophysiology, Center of Internal Medicine, Univ. of Essen, School of Medicine, Hufelandstrasse 55, 45122 Essen, Germany.

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract